Organic Process Research & Development 2003, 7, 109−114
Solid/Liquid Palladium-Catalyzed Coupling of Haloaryls Using Alcohols as Reducing Agents: Kinetics and Process Optimization Diana Gitis,† Sudip Mukhopadhyay,‡,| Gadi Rothenberg,§,⊥ and Yoel Sasson*,† Casali Institute of Applied Chemistry, Hebrew UniVersity of Jerusalem, 91904, Israel, Chemical Engineering Department, UniVersity of California at Berkeley, Berkeley, California 94720, U.S.A., and Chemical Engineering Department, UniVersiteit Van Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands
Abstract: Heterogeneous Pd-catalyzed coupling reaction of substituted halobenzenes to biphenyls is performed effectively in different organic solvents using an alcohol as the reducing agent in the presence of a base and a catalytic amount of phase-transfer (PT) agent. The alcohol is oxidized to the respective aldehyde or ketone. Kinetic studies show that the reaction rate equation fits to second order in halobenzene. The effects of reaction temperature, catalyst loading, base concentration, PTC concentration, and alcohol concentration and structure are studied and discussed.
Introduction Industrial interest in the synthesis of biaryl compounds reflects their expanding role in natural product synthesis and polymer chemistry, and their usage as key building blocks for many agrochemicals and fine-chemical intermediates.1 One of the widespread methods in biaryl synthesis is the Ullmann reaction.2 This method is effective but uses an excess of copper.3 Several catalytic alternatives exist such as the Suzuki,4 Stille,5 Negishi couplings. Aryl-aryl coupling via a radical mechanism under photochemical conditions6 and intramolecular radical arylations under reductive radical conditions using tributyltin7 have also been reported. Those reactions are possible due to formation of intermediates such as radical anions of haloaromatic compounds. In several cases these reactions take place in the presence of palladium catalysts.8 These palladium-catalyzed reductive coupling reactions involve the interconversion Pd(0)/Pd(II), that is between d10 * Corresponding author. Telephone: +972-2-658-4530. Fax: +972-2-6528250. E-mail
[email protected]. † Hebrew University of Jerusalem. ‡ University of California at Berkeley. § Universiteit van Amsterdam. | E-mail:
[email protected]. ⊥ E-mail:
[email protected]. (1) (a) Stanforth, S. P. Tetrahedron 1998, 54, 263. (b) Bringmann, G.; Walter, R.; Weirich, R. Angew. Chem., Int. Ed. Engl. 1990, 29, 977. (c) Sainsbury, M.; Tetrahedron 1980, 36, 3327. (2) Ullmann, F. Ber. 1903, 36, 2389. (3) Stinson, S. C. Chem. Eng. News 1999, 77, 65. (4) (a) Suzuki, A. Pure Appl. Chem. 1991, 63, 419. (b) Suzuki, A. J. Organomet. Chem. 1999, 576, 147. (5) de Meijere, A.; Meyer F. E. Angew. Chem., Int. Ed. Engl. 1995, 33, 2379. (6) Taylor, E. C.; Kienzie, F.; McKillop, A. J. Am. Chem. Soc. 1970, 92, 608. (7) Clive, D. L. J.; Kang, S. Tetrahedron Lett. 2000, 41, 1315. (8) Tsuji, J. Palladium Reagents and Catalysts 1996; Wiley: Chichester. See also (b) Mukhopadhyay, S.; Rothenberg, G.; Lando, G.; Agbaria, K.; Kazanci, M.; Sasson Y. AdV. Synth. Catal. 2001, 343, 455. 10.1021/op015519+ CCC: $25.00 © 2003 American Chemical Society Published on Web 12/19/2002
and d8 electronic configurations. In the case of heterogeneous Pd/C, completion of the catalytic cycle via reduction of the palladium catalyst was achieved using hydrogen gas (Scheme 1),9 dihydrogen-generating agents (e.g., formate/water10 or zinc/water11), and redox H-transfer agents (e.g., benzene12 or hydroquinone13). Previously we suggested that the catalytic cycle involves two single-electron-transfer (SET) processes from the Pd(0) catalyst to two haloaryl molecules,14 followed by dissociation of the [Ar-X]•- radical anions.15 Scheme 1. Reductive coupling of halobezenes in the presence of a palladium catalyst
The main drawbacks of these reducing agents are the competitive chemisorption and hydride-forming ability, which lead to the reductive hydrodehalogenation of aryl halides. Accordingly, one might envisage that a mild reducing agent (e.g., an alcohol) would minimize this side reaction.16 Here we describe studies on the coupling reactions of bromoaryls using alcohols as the reducing agent under basic conditions and discuss some of the elementary processes in this system. (9) Mukhopadhyay, S.; Rothenberg, G.; Wiener, H.; Sasson, Y. Tetrahedron 1999, 55, 14763. (10) Mukhopadhyay, S.; Rothenberg, G.; Gitis, D.; Sasson, Y. J. Org. Chem. 2000, 65, 3107. (11) Mukhopadhyay, S.; Rothenberg, G.; Gitis, D.; Sasson, Y. Org. Lett. 2000, 2, 211. (12) Hennings, D. D.; Iwama, T.; Rawal, V. H. Org. Lett. 1999, 1, 1205. (13) Tamura, Y.; Chun, M.-W.; Inoue, K.; Minamikawa, J. Synthesis 1978, 822. (14) Mukhopadhyay, S.; Rothenberg, G.; Gitis, D.; Wiener, H.; Sasson, Y. J. Chem. Soc., Perkin Trans. 1999, 2481. (15) For a comprehensive review, see: (a) Dalko, P. I. Tetrahedron 1995, 51, 7579. See also: (b) Munavalli, S.; Rossman, D. I.; Szafraniec, L. L.; Beaudry, W. T.; Rohrbaugh, D. K.; Ferguson, C. P.; Gratzel, M. J. Fluor. Chem. 1995, 73, 1. For studies on the electron-transfer rate constants for C-X bond cleavage see: (c) Save´ant, J. M. Acc. Chem. Res. 1993, 26, 455. (d) Save´ant, J. M. AdV. Phys. Org. Chem. 1990, 26, 1. (16) The application of alcohols in the Pd-catalyzed coupling of aryl halides was first described by Mayo et al., see Mayo, F. R.; Hurwitz, M. D. J. Am. Chem. Soc. 1949, 71, 1. Vol. 7, No. 1, 2003 / Organic Process Research & Development
•
109
Table 1. Coupling of various starting materialsa entry
substrate
1 2 3 4
bromobenzene 4-bromotoluene 4-bromoanisole 4-chloro-1,1,1,-trifluorotoluene
conversion selectivity to PhX/%b coupling/%b 92.0 100.0 91.2 78.6
61.0 65.1 47.8 58.0
a Standard reaction conditions: 32 mmol PhBr, 64 mmol NaOH, 35 mmol 2-pentanol, 1.6 mmol 5% Pd/C (5 mol % of Pd relative to PhBr), 3.2 mmol PEG-400, 105 °C, 25 mL of toluene (solvent), 6 h. b Based on GC area, corrected by the presence of an internal standard.
Results and Discussion In a typical reaction, one molar equivalent of halobenzene was stirred in xylene with 1.1 equiv of alcohol, in the presence of 2 equiv of NaOH and catalytic amounts of Pd/C and (poly)ethyleneglycol (PEG). The reaction yielded mainly biphenyl along with some benzene (eqs 1 and 2), and the
Figure 1. Coupling reaction profile. Conditions: 32 mmol PhBr, 64 mmol NaOH, 35 mmol 2-pentanol, 1.6 mmol 5% Pd/ C, 3.2 mmol PEG-600, 105 °C, 25 g of xylene (solvent).
Figure 2. Determination of the reaction order XA ) fractional conversion of bromobenzene defined as XA ) 1 - NA/NAO (conditions are as in Figure 1).
alcohol was oxidized to the corresponding aldehyde/ketone. Good yields of the coupling products were obtained using various substrates as shown in Table 1. No other products were observed by GC and 1H NMR analysis. No coupling took place in the absence of any one of the sodium hydroxide, alcohol, or catalyst. PhBr was chosen as a model substrate for the process parameter studies. A typical profile for the coupling reaction of PhBr is shown in Figure 1. Material balance showed complete conversion of PhBr: 60% to biphenyl and 40% to benzene. The reaction was found to fit a second-order profile (Figure 2). The reaction temperature was varied from 90 to 125 °C. As expected, higher temperature increased the reaction rate but did not much affect the product selectivity (Table 2, entries 1-4). Reaction rates increased from 0.9 × 10-4 to 25 × 10-4 s-1 M-1. The Arrhenius energy of activation was found to be Ea ) 110 kJ mol-1 (26.3 kcal mol-1, with r2 ) 0.974 for four measurements at 90, 105, 115, and 125 °C, see Figure 3). This relatively high Ea value indicates that the rate-determining step is chemically controlled. The catalyst loading, expressed as the wt % of catalyst based on substrate concentration, was varied from zero to 110
•
Vol. 7, No. 1, 2003 / Organic Process Research & Development
10.0% (Table 2, entries 5-12). At low loadings ( cyclohexanol > 2-pentanol > 1-pentanol > tert-butyl alcohol > tert-amyl alcohol. Benzylic alcohols are the best hydrogen donors, probably due to their high affinity for the metal. Conversely, the R-hydrogen of a primary alcohol is less likely to react as a hydride species because of the smaller electron-releasing inductive effect of the alkyl group.18 Note that tert-butyl alcohol and tert-amyl alcohol were also reactive, showing that the presence of R-hydrogen on the alcohol is not critical (and obviously that the above pathway is not the only one possible). Although the reaction rate was highest with benzyl alcohol (kobs ) 8 × 10-4 M-1 s-1), the selectivity to biphenyl was not satisfactory. Secondary alcohols were found to be the preferred Ar-Ar coupling reagents. Changing the alcohol concentration did not influence the reaction rate, but product selectivity proved sensitive to the amount of alcohol (increasing the latter decreased the selectivity to biphenyl due to a higher dehydrohalogenation (17) Ukisu Y.; Miyadera T. J. Mol. Catal. A 1997, 125, 135. (18) Besson, M.; Gallezot, P. Catal. Today 2000, 57, 127.
Table 2. PhBr coupling under various conditionsa entry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
parameter changed t/°C 90 105 115 125 5% Pd/C /mol % relative to PhBr 0.0 0.5 1.0 1.5 2.5 5.0 7.5 10.0 KOH /mol % relative to PhBr 0 30 100 200 400 PEG-600 /mol % relative to PhBr 0.0 2.5 5.0 10.0 20.0 alcohol /mol % relative to PhBr 0 25 50 110 200
time/h
conversion PhBr/%b
selectivity biphenyl/%b
2 2 2 2
24.0 72.0 77.0 85.0
51.0 64.0 54.0 61.0
6 6 6 6 6 6 6 6
0.0 17.0 22.8 26.1 75.0 92.0 77.0 82.5
0.0 11.0 12.2 19.7 64.0 61.0 59.0 51.2
3 3 3 3 3
0.0 14.0 52.0 94.0 88.0
0.0 80.0 68.0 56.0 27.0
4 4 4 4 4
5.0 16.0 37.0 81.0 100.0
78.0 47.0 63.0 62.0 57.0
20 20 20 20 20
0.0 51.0 72.0 100.0 100.0
0.0 70.0 62.0 60.0 48.0
a Standard reaction conditions: 32 mmol PhBr, 64 mmol NaOH, 35 mmol 2-pentanol, 1.6 mmol 5% Pd/C (5 mol % Pd relative to PhBr), 3.2 mmol PEG-400, 105 °C, 25 mL of toluene (solvent), 6 h. b Based on GC area, corrected by the presence of an internal standard. c Cetyl-trimethylammonium bromide (CTAB) was used instead of PEG.
Figure 3. Arrhenius plot. Conditions: 32 mmol PhBr, 64 mmol NaOH, 35 mmol 2-pentanol, 1.6 mmol 5% Pd/C, 3.2 mmol PEG-600, 25 g of xylene (solvent).
rate). Note that PEG also acted as the terminal reductant of PdII (Table 2, entries 23-27). Type of Base. The rate of conversion was found to be dependent on concentration of the base (e.g., on [KOH], as shown in Figure 4). In the absence of base, neither reduction nor coupling occurred. Changing the concentration of KOH from 30 to 200 mol % increases the rate 12-fold (from 0.5 × 10-4 M-1 s-1 to 2.3 × 10-3 M-1 s-1). This is in agreement with previous studies that showed that reaction velocity and the attainable concentrations of radical anions increased at higher [OH-].19 The role of base compounds is not limited to neutralization of HBr.20 The base probably promotes both
hydrogen-atom-transfer and the electron-transfer reactions.21 Adding less than two equivalents of KOH resulted in incomplete conversion. The concentration of base affects the selectivity (Table 2, entries 13-17). A maximum of 80% selectivity occurs with a small 30% concentration of base. Further increase in base concentration decreases the selectivity down from 80 to 30%, and these values were essentially independent of time of reaction. Weaker bases (K2CO3, NaHCO3, and Et3N) were also examined, but conversions were ArBr > ArCl.26 However the selectivity of these reactions to biphenyl does not show the same trend. The initial electron transfer and formation of radical anions from halobenzenes is expected to be a function of the strength of the Ar-X bond.27 Our results may arise from the combination of two opposite (26) As expected, no reaction was observed with fluorobenzene. (27) Pierini A. B.; Duca J. S. J. Chem. Soc., Perkin Trans. 2 1995, 1821.
Figure 6. Effect of different initial concentrations of PhBr on initial reaction rate. Conditions: 64 mmol NaOH, 35 mmol 2-pentanol, 1.6 mmol 5% Pd/C, 3.2 mmol PEG-600, 105 °C, 25 g of xylene (solvent).
factors: (i) the C-X bond energies and (ii) the adsorption strength of the halobenzene on the catalyst. The former decrease in the order DC-F (453.6 kJ mol-1) > DC-Cl (340.2 kJ mol-1) > DC-Br (281.4 kJ mol-1) > DC-I (222.6 kJ mol-1).18 However, the adsorption strength of aryl halides decreases in the following order: PhI . PhCl ≈ PhBr.28,29 The high reaction rate and good selectivity observed for PhBr may be therefore due to the combination of moderate adsorption strength and relatively low C-Br bond energy, supporting the suggestion that the rate-determining step in the reaction is the adsorption of the halobenzene on the Pd surface (followed by electron transfer).30 The low selectivity to biphenyl observed in the case of PhI reflects its high adsorption strength on Pd catalyst. Mechanistic Considerations. When all other reagents are used in excess, the rate of the coupling reaction (eq 1) shows second-order dependence on [PhBr], that is r1 ) k1[ArBr]2. Likewise, we can express the rate of the reduction reaction (eq 2) as r2 ) k2[ArBr], making the overall rate r ) k1[ArBr]2 + k2[ArBr]. When k1 . k2, the rate reduces to the secondorder, fitting with the experimental observations (Figure 2 above). Figure 6 shows the influence of different initial concentrations of PhBr on the initial reaction rate. An optimum PhBr concentration is observed, above which the reaction rate decreases. The shape of this curve is consistent with the Langmuir-Hinshelwood model pertaining to adsorption effects on a heterogeneous catalyst when large quantities of starting material cover the Pd surface, so that other reagents are denied access to the surface.31 Halobenzenes are known to react with various metals to form radical anions.32 Diffusion of these radical anions from the surface of the metal into the solution frequently leads to radical coupling/reduction products.33 In our system, the radical anion formed from PhBr in the presence of Pd/C (28) Aramendia M. A.; Borau V.; Garcia I. M.; Jimenez C.; Marinas A.; Marinas J. M.; Urbano F. J. Surf. Chem. Catal. 2000, 3, 465. (29) The relative adsorption coefficients are DPhBr/PhCl ) 0.88, DPhI/PhCl ) 497, and DPhI/PhBr ) 564. See Rader C. P.; Smith H. A. J. Am. Chem. Soc. 1962, 84, 1443. (30) Bodewitz, H. W. H. J.; Blomberg, C.; Bickelhaupt, F. Tetrahedron 1973, 29, 719. (31) Aramendia, M. A.; Borau, V.; Garcia, I. M.; Jimenez, C.; Marinas, J. M.; Urbano, F. J. Appl. Catal. 1999, B 20, 101. (32) (a) Tsou, T. T.; Kochi, J. K. J. Am. Chem. Soc. 1979, 101, 6319. (b) Zask, A.; Helquist, P. J. Org. Chem. 1978, 43, 1619. (33) Orchin M. J. Chem. Ed. 1989, 66, 586. Vol. 7, No. 1, 2003 / Organic Process Research & Development
•
113
yields a mixture of biphenyl and benzene. The mechanism of the present catalytic reaction seems intriguing as it is different from the well-known coupling reaction in the presence of formate salts. Further studies on the scope and mechanism of this catalytic system are in progress. Conclusions In the absence of water the stoichiometric reductive coupling of halobenzenes was shown to have very low activity. However, this reaction can be achieved using alcohols. Alcohols can be utilized as hydrogen-transfer agents for the coupling of halobenzene in the presence of heterogeneous Pd/C. The reaction was found to be sensitive to the type and concentration of the base. The absorbance strength of ArX on the heterogeneous palladium catalyst and the nature of the support play a key role in the coupling and reduction of halobenzenes. Experimental Section Materials, analytical methods, and instrumentation have been described in detail in a previous publication.14 General Procedure for Haloaryl Coupling. Example: Biphenyl from PhBr: 5.0 g (32 mmol) of PhBr, 3.1 g (35 mmol) of 2-pentanol, 3.3 g of 5% Pd/C (5 mol % Pd based on PhBr), 2.56 g (64 mmol) of NaOH, 1.92 g of PEG-600 (10 mol % based on PhBr), 0.8 g of naphthalene (internal standard), and xylene (solvent, total reaction volume 40 mL) were charged to a flask that was heated to 105 °C. Reaction progress was monitored by GC. The mixture was stirred for 8 h, after which 91.2% conversion of PhBr was measured. Yields: biphenyl 62% (1.38 g), benzene 38% (0.86 g). Kinetic Studies. The standard reaction conditions were as in the general procedure. The following parameters were studied: (i) initial PhBr concentration, (4 experiments at 16 mmol, kobs ) 9 × 10-4 s-1 M-1, r2 ) 0.983 for five observations; 32 mmol, kobs ) 3 × 10-4 s-1 M-1, r2 ) 0.990 for five observations; 48 mmol, kobs ) 1 × 10-4 s-1 M-1, r2 ) 0.981 for five observations; 64 mmol, kobs ) 0.8 × 10-4 s-1 M-1, r2 ) 0.975 for 5 observations); (ii) catalyst loading (six experiments using 0.5 mol % of 5% w/w Pd/C, kobs ) 0.4 × 10-4 s-1 M-1, r2 ) 0.979 for five observations; 1.5
114
•
Vol. 7, No. 1, 2003 / Organic Process Research & Development
mol %, kobs ) 0.5 × 10-4 s-1 M-1, r2 ) 0.951 for five observations; 2.5 mol %, kobs ) 2 × 10-4 s-1 M-1, r2 ) 0.993 for five observations; 5.0 mol %, kobs ) 3 × 10-4 s-1 M-1, r2 ) 0.983 for five observations; 7.5 mol %, kobs ) 2 × 10-4 s-1 M-1, r2 ) 0.985 for five observations; and 10.0 mol %, kobs ) 1 × 10-4 s-1 M-1, r2 ) 0.99 for five observations); (iii) PTC loading (four experiments using 2.5 mol % PEG-600, kobs ) 0.1 × 10-4 s-1 M-1, r2 ) 0.999 for five observations; 5 mol % PEG-600, kobs ) 0.5 × 10-4 s-1 M-1, r2 ) 0.999 for five observations; 10 mol % PEG-600, kobs ) 3 × 10-4 s-1 M-1, r2 ) 0.993 for five observations; and 20 mol % PEG-600, kobs ) 28 × 10-4 s-1 M-1, r2 ) 0.998 for five observations); (iv) reaction temperature (four experiments at 90 °C, kobs ) 0.9 × 10-4 s-1 M-1, r2 ) 0.993 for five observations; 105 °C, kobs ) 3 × 10-4 s-1 M-1, r2 ) 0.999 for five observations; 115 °C, kobs ) 9 × 10-4 s-1 M-1, r2 ) 0.989 for five observations; 125 °C, kobs ) 25 × 10-4 s-1 M-1, r2 ) 0.998 for five observations); (v) KOH loading (four experiments using 30 mol % of KOH, kobs ) 0.5 × 10-4 s-1 M-1, r2 ) 0.972 for five observations; 100 mol % of KOH, kobs ) 3 × 10-4 s-1 M-1, r2 ) 0.956 for five observations; 200 mol % of KOH, kobs ) 2.3 × 10-3 s-1 M-1, r2 ) 0.994 for five observations; 400 mol % of KOH, kobs ) 5 × 10-4 s-1 M-1, r2 ) 0.996 for five observations); (vi) alcohol (2-pentanol) concentration (four experiments using 25 mol % of 2-pentanol, kobs ) 1 × 10-4 s-1 M-1, r2 ) 0.996 for five observations; 50 mol %, kobs ) 2 × 10-4 s-1 M-1, r2 ) 0.997 for four observations; 100 mol %, kobs ) 3 × 10-4 s-1 M-1, r2 ) 0.975 for five observations; and 200 mol %, kobs ) 3 × 10-4 s-1 M-1, r2 ) 0.980 for five observations. Acknowledgment We thank the Israel Ministry of Science for an Eshkol Individual Fellowship (to D.G.). The invaluable support and keen interest of Professor M. Orchin (Department of Chemistry, University of Cincinnati) is deeply appreciated. Received for review December 7, 2001. OP015519+
